Computer science enrolment falls 25%

Where things stand (from the Clearinghouse's own data files)

I downloaded both Clearinghouse major-field appendices and added up the three undergraduate sectors for Computer and Information Sciences (CIP 11): 4-year, PAB and 2-year.

• Fall 2024 baseline (B): 950,702 (659,708 + 54,332 + 236,662), as restated in the Final Fall Enrollment Trends 2025 appendix (xlsx, Jan 15 2026). The threshold, 0.75×B, is 713,026.
• Fall 2025: 875,459, down 7.9% year on year (4-year −8.1%, PAB −3.6%, 2-year −8.3%).
• Spring 2026: 818,714, down 9.2% year on year (4-year −8.4%, PAB −9.3%, 2-year −11.2%). That is 86.1% of B (xlsx, June 4 2026). This release predates the question's opening date, so it cannot resolve the question.
• Revisions are small. Fall 2023 was first published as 910,159 and later restated as 918,354 (+0.9%). Spring 2025 went from 900,117 to 901,320 (+0.1%).
• Spring runs below the preceding fall. Spring as a share of the prior fall was about 0.98 in the growth years, then 0.948 (2024–25) and 0.935 (2025–26). Because the question allows spring-to-fall comparisons, this seasonal dip lowers the bar for spring terms.

What YES requires

• Fall 2026 data (January 2027 report): a year-on-year fall of at least 18.6%.
• Spring 2027 data (June 2027 report): a year-on-year spring fall of at least 12.9%. That roughly means fall 2026 down 13% or more.
• Fall 2027 or spring 2028: about 10% a year, or about 7% a year, over two years.

Release calendar. The Clearinghouse's homepage lists the Preliminary Fall report for late October 2026; it is panel-based and does not qualify. The Final Fall report comes in January and the Final Spring report in early June. Only the first and second quarters of each year can therefore move the forecast.

Reference class and dynamics

• Past computer science busts were deep and long. In the dot-com bust, Computing Research Association data showed undergraduate enrollment in PhD-granting departments down 39% from its peak after four years and 49% after five. Annual drops in later years reached 14–18% (CRA 2007; Webster 2008).
• Stock and flow. Undergraduate computer science credentials hit a record 210,100 in 2024–25 (NSC Degree Earners, Apr 2026). The large classes that entered in fall 2022 and 2023 graduate in 2026–27 while fewer students enter. My rough model gives about −8% to −9% in fall 2026 even with flat inflow, and −11% to −14% if inflow falls another 10–20%.
• Leading indicators. New majors fell 13% at PhD-granting departments while their total enrollment fell 4% (CRA Taulbee, Jun 2026). In a fall 2025 survey, 62% of computing departments reported declines (CRA, Oct 2025). University of California computer science majors fell 3% in 2024 and 6% in 2025 (GovTech/SF Chronicle, Feb 2026). Ontario computing applications were down 44% from their 2023 peak in the 2026 cycle (Globe and Mail, Sep 2026). Coverage of a weak job market for new graduates continued through September 2026 (AP).

Model

I ran a Monte Carlo simulation with these assumptions:

• Fall 2026 year-on-year change is a mixture: 78% "continued decline" centred at −12% (SD 3.3 points) and 22% "stabilisation" centred at −8.5%, with recovery afterwards.
• Later years follow with persistence and gradual easing.
• Spring as a share of the prior fall is about 0.959 + 0.2 × the next fall's change.
• Small noise for baseline restatements, plus about 1% a year chance that the data series stops being usable.

Results:

• Fall 2026 crossing in the January 2027 release: about 1–2%.
• Spring 2027 crossing: about 25–32%. In the central case spring 2027 lands near 76–77% of B, just above the line.
• By January 2028: about 50–58%. By June 2028: about 70–76%.
• Plateau by 2031: about 0.79–0.82. I shaded this to about 0.77 for structural uncertainty.

Strongest case against

The 4-year sector, 69% of the total, has not clearly accelerated (−8.1% in fall, −8.4% in spring). IT and cybersecurity programs (CIP 11.10), about 20–25% of the field, are roughly flat. Selective programs with capped seats absorb weaker demand without shrinking. AI majors that are coded within CIP 11 may grow. If inflow stabilises soon, the decline could level off at about −15% to −20% and never cross. I give this roughly a 20–25% weight.

The case the other way is that crossing only needs about −20% in fall terms once the spring dip is used, and the large graduating cohorts keep pushing enrollment down through 2027 and 2028.

Reading adopted

For spring terms I compare against fall 2024 as restated in the latest January release. Preliminary panel results are treated as not qualifying.

Question. By the last day of each quarter, has the National Student Clearinghouse reported U.S. undergraduate computer science enrolment at least 25% below its fall 2024 level?

Definitions.

• National Student Clearinghouse: the National Student Clearinghouse or its Research Center (NSCRC).
• Computer science enrolment: undergraduate enrolment in the NSCRC major family Computer and Information Sciences (Classification of Instructional Programs, two-digit CIP 11), not only the narrower Computer Science subfield. Use enrolment counts, not full-time-equivalent enrolment, credit hours, course registrations, applications, entrants, degrees awarded, or the field's share of all enrolment.
• Coverage: the national total across all institutional sectors and undergraduate award levels included in NSCRC's major-field statistics, including four-year institutions, primarily associate-degree-granting baccalaureate institutions, and two-year institutions. Graduate enrolment is excluded. Use NSCRC's counting and coverage-adjustment methodology; do not independently deduplicate its counts. If no aggregate is provided, sum mutually exclusive undergraduate institutional groups covering the full reported population.
• Comparison: let B be the reported fall 2024 count and E the count for a subsequent regular fall or spring term. The threshold is E ≤ 0.75 × B. Spring-to-fall comparisons count; summer-only counts do not. Both counts must use comparable field definitions, populations, and methodology.
• Baseline and revisions: use fall 2024 as restated in the same data release as E, if available; otherwise use the most recent comparable official fall 2024 figure available when E is published. The original baseline source is Current Term Enrollment Estimates: Fall 2024 — Major Fields Data Appendix, published January 23, 2025. A reclassification or methodology change counts only where NSCRC provides a comparable fall 2024 baseline or an explicit like-for-like comparison.

Resolves YES if an eligible NSCRC publication provides counts satisfying the threshold, or explicitly reports a decline of at least 25% from fall 2024 for this complete population. A calculation from official counts is sufficient; NSCRC need not state the threshold in words. Use the most precise published figures. If counts are unavailable, an explicitly reported decline of 25.0% or greater qualifies. Do not compound rounded annual percentage changes to infer an otherwise unreported cumulative decline.

The Resolution window and Observation period conventions apply: the qualifying publication must be dated on or after opening, and the measured term must end after opening. A later correction does not undo a qualifying YES.

Resolves NO if no qualifying publication is available by the quarter-end. Declines confined to four-year colleges, bachelor's programmes, freshmen, a narrower CIP subfield, or another subset do not qualify. A preliminary fixed-panel comparison alone does not establish the national total; preliminary results qualify only if they provide comparable national enrolment estimates or an explicit national comparison meeting the definitions above. No attribution of the decline to AI is required.

Resolving source. NSC/NSCRC-authored reports, data appendices, dashboards, or official statements on nscresearchcenter.org or studentclearinghouse.org only, a narrower source rule than Public confirmation (A2). Primary sources are Clearinghouse Enrollment Insights, especially Final Fall Enrollment Trends and Final Spring Enrollment Trends, their Major Fields Data Appendices, and the fall 2024 appendix. Official successor publications count if they preserve comparability. Apply Resolution date (A3) to the publication or dated data update, not the enrolment term.